The hypothesis that occupational noise exposure is positively associated with hypertension was examined in a cross-sectional study carried out on a group of patients who were enrolled at the Occupational Health Unit of the Unified Health System, situated in Salvador city, the capital of Bahia state, Brazil. Data were obtained from 276 medical records, corresponding to all patients newly registered during the first six months of 1992. Data on noise exposure come from both reported occupational exposure history and clinical diagnosis of occupational noise-induced hearing loss. Hypertension diagnosis complies with World Health Organization criteria, as well as with the history of antihypertensive treatment. Stratified analysis and unconditional logistic regression modeling show results that do not support the study hypothesis: there are no differences between systolic or diastolic blood pressure or between proportion of hypertension for exposed and non exposed groups. However, statiscally significant (alpha = 0.05) increment of the effect measured was reported among workers who reported low educational level (below elementary). This could be another evidence of socially related inequalities underlying exposure distribution among workers at the workplace, which should be addressed, at greater depth, in future studies.